Hello there Jesper.
Well this is a big field, and spell fixes.. well it depends so much on how the spell is broken. Sometimes a simple database fix for it is working out, while some spells may need to get hardcoded in the core to work properly - it depends so much on what function it should have and what it currently does etc.
For a basic understanding you should defnieatly check out
http://collab.kpsn.org/pages/viewpage.action?pageId=10682946 where you can get some more information.
I'm going to like you a few other TC sources where you can read and hopefully understand some of it which should make the situation easier,
if you are really interested in learning it.
http://collab.kpsn.org/display/tc/Spell+system - including coding examples & structure.
http://collab.kpsn.org/display/tc/Spell+handling+in+TC
And some spell flags definitions that may be useful -
http://collab.kpsn.org/display/tc/Spell+Effects+Reference
Here's some basic database knowledge as well:
http://collab.kpsn.org/display/tc/SQL+Guidelines
Again, some spells are able to get fixed in the database, this could for example be in spell_proc_event if the spell misses a simple proc or what so ever, could also get hack fixed in spell_linked_spell so that you set the original spell to use a trigger from another spell that for example works.
You should defineatly check out some talents etc and see how they're fixed.
The default places the spell fixes are added is typically in spell_class.cpp (spell_mage.cpp etc), unit.cpp, spellauraeffects and spelleffects.cpp but again it's huge and there's still a lot to know and you may want to learn your way in the core.
I hope this was helpful somehow.
